Infrared internal reflection spectra of methanol—water mixtures

Infrared internal reflection spectra of methanol—water mixtures

Spwtmotdmicadata. Vol. SOA. pp. 1787 to 1792. PergrrmonPrea 1074. Printed in Northern Ireland InfraredinternalMb&ion spectraof methanol-watermihres R...

392KB Sizes 0 Downloads 15 Views

Spwtmotdmicadata. Vol. SOA. pp. 1787 to 1792. PergrrmonPrea 1074. Printed in Northern Ireland

InfraredinternalMb&ion spectraof methanol-watermihres R. T. Yma

and M. J. D. Low*

Department of Chemistry, New York University, 4 Washington PIsee, New York, N.Y. 10003, U.S.A. (Rec&ued18 00th

1973)

AbotrrraeInfraredeptmtraof O-100 ~01% methanol-water mixtures were reoorded over the

3900-900 cm-l range at 8 om-’ resolution using internal reflection teohniques and a Fourier transform spectrometer. There were inters&ions of the water and methanol which made themselves evident as changes of the water OH stremhing and deform&ion band frequencies. The methanol band frequerioieswere unsffeoted by ohanges in coneentrstion and oonsequently those bands can be used for quantitative purposes, but weter bands should not be used for precise determinations.

M&oxlo

and B'LOIJBNOY have measured infrared intern&l reflection spectra of of methanol, ethanol, n-propmol, and i-propanol, and reported that Beer’s Law concentration dependencies were obtained over O-100% concentration ranges, so that the measurements would be useful as a quantitative analytical procedure [l]. No band frequency changes were observed, except for the skeletal bsnds of n-proptmol, and they commented that “ . . . no explanation has been derived for the surprisingly linear relationships found for these non-ideal solutions” [l]. As they pointed out for the case of methanol-water mixtures, the absence of detectable shifts in vibrational frequencies over the wide range of concentrations indicates that there were no significant changes in intermolecular bonding. However, these general observations conflict with earlier ones made with ethanol-water mixtures: SZEPESYet al., recorded transmission spectra of ethanolwater mixtures in the overtone region and reported finding Beer’s Law deviations which they attributed to ethanol-water interactions [2]. Also, SPINKand WYOXOFF measured near-infrared differential spectra of several simple alcohols in aqueous solution and found chsnges in the 860-1020 nm region which they used to derive the appsrent hydration numbers of the alcohols [3]. This discrepancy w&s of interest to us because changes in solute-solvent interactions are pertinent to our infmred studies of adsorption phenomena ocourring st the liquid-solid interface [a]. As we contemplated using adsorb&s dissolved in methanol-water mixtures, the methanol-water system was reinvestigated. aqueous solutions

EXPERIMENTAL Internal reflection techniques were employed using a Ge prism in a suitable attachment [a]. The incident angle was adjusted to give an equivalent penetration * To whom inquiries should be addressed. [I] C. P. -ONE and P. A. FLOUXNOY, Spectmchim. Acta 81, 1361 (1966). [2] G. L. SZEPESY,J. CSASZAR,and L. h!EfOTa, Acta Univ. Szegedienstk, Acta Phyu. et Chim. 2, 149 (1966). I31 C. H. SPIN-Kand J. C. Wyc~olra, J. Phys. Chem. 76, 1660 (1972). [4] R. T. YANG, M. J. D. LOW, G. L. IXALLER, and J. FENN, J. COLLOID,IraterfaceSci. 4,249

(1973).

1787

R. T. YANU and M. J. D. Low

1788

depth of 0.86 pm at 3300 cm-l. Spectra having a constant resolution of 8 cm-1 were recorded with a modified Digilab, Inc. Model FTS-14 Fourier transform spectrometer [KJ. All single-beam spectra were obtained by summing 400 interferograms. After the cell containing the prism was installed, the sampling compartment was flushed with dry air. The cell remained in place throughout the series of measurements so that its transmission remained constant; small changes in cell position can cause transmission changes which can affect the results obtained when two single-beam spectra are ratioed. Aqueous solutions of 0, 10,26,60,75 and 100 ~01% methanol, prepared with spectral-grade absolute methanol and distilled water, were placed into the cell or withdrawn from it by means of syringes and Teflon tubing. Solutions of reagent-grade NaClO, in methanol were similarly treated. RESETS

AND D~scnss~o~

Single-beam spectra were stored, recalled, and ratioed by computation as required, using the instrument’s digital computer. Some examples are shown in Fig. 1. The ratioing process is equivalent to using a dispersion spectrometer in the

3000

2000

IOU0 cm”

Fig. 1. Single-beam spectra: A: dry cell. B: pure methanol. in water. D: pure water.

c:

60 % methanol

double-beam mode with cells of precisely the same optical quality and path lengths in each beam, thus “compensating” one liquid with another. This equivalency is correct only when the two liquids whose spectra are ratioed have identical refractive indices, so that the equivalent penetration depths for the two liquids are the same. In the present case, with the indices of refraction at 26’ of water and methanol of 1.332 and 1.326, respectively, the penetration depths differ by less than Oil%. Also, the index matching is good, the high refractive index of the Ce prism causing any [S] M. J. D. Low, A. J.

GOODSEL, and H. MARE, in Molecoclar Spectroscopy (1971), (edited by EiehComf. Molecular Spectroscopy, Brighton, England, 1971. Institute of Petroleum, London, 1972; pp. 383ff.

HEPPLE P.) Proc.

Infrared internal refleationspectra of methanol-water mixtures

1789

effects due to variation of the refractive index of the sample to be very small. At low absorptivity the variation of the refractive index in the absorption band is also negligible. Consequently, possible errors caused by these effects can be negleoted. The single-beam spectra of the pure liquids or solutions can be ratioed against “air,” i.e., against the spe&um of the empty cell, to obtain spectra such as A, B, or C of Fig. 2, in which the instrument function has been removed. The contour of the

h.,....,.........,.........,l

3000

2000

1000 err?

Fig. 2. Ratioed speotra: A: water ratioed against air (the ratio of spectra D and A of Fig. 1). B: methanol ratioed against air (the ratio of spectra B and A of Fig. 1). C: 60 ‘A methanol solution ratioed against air (the ratio of spectra C and A of Fig. 1). D: 75% methanol solution ratioed against methanol (the ratio of the single-beam spectrum of a 76 % methanol solution similar to spectrum C of Fig. 1 against spectrum B of Fig. 1).

OH stretching band (A, Fig. 2) is like that of the infrared transmission spectrum of a thin water layer reported by THOMPSON et al, [6], who concluded that there were three strongly overlapping bands at 3625, 3410 and 3260 cm-l. The OH band (A, Fig. 2) is also similar to the Raman OH band [7], but differs from that of BONNER and CURRY [8, 91. The spectra of water, methanol, and water-methanol solution [6] [7] [S] [9]

W. K. THOD~P~ON, W. A. SENIOR,and B. A. PETIXICA, Nature, Lo&. 211,1080 (1966). W. F. MELRPIXY and H. J. BERNSTEIN, J. Phy8. C%em.76, 1147 (1972). 0. D. BONNERand J. D. CURRY,IrzfmredP&&a 10, 91 (1970). BONNERand CURRY[S] recorded differentialspectra of water using two cells of nominally 0.007 mm pathlength but differing in actual pathlength by 6-&O%, and found two quite definite bands at 3630 and 3200 cm-’ (2625 and 2418 cm-l with DaO). Their “effective thickness” of liquid yielding their differentialspectra was of the order of O-0006 mm, i.e., close to the effective penetration depth of 0@0085 mm in the present work. However, the actual thickness of their liquid was near O-007 mm, and a single layer of it had a trsnsmittance close to zero near 3400 cm- r. The cause of the difIerenoesin the OH bands is not known. Possibly, the change in the refractive index in the region of intense absorption caused changes in the absorptive, refractive, and reflectiveproperties of the thin films; or possibly, the relatively fast scan rates used in regions of high absorption caused some distortion.

1790

R. T. YANQ and M. J. D. Low

(A, B, C, Fig. 2) are thus similar to those of -NE and FLOURNOY (Fig. 2 of reference l), who had an attachment with a &liquid interface in the sample beam ctnd an identical attachment with a Ge-air interface in the reference beam [ 11. When the single-beam spectra are ratioed against those of the pure solvents, solvent-compensated spectra such as those of Fig. 3 are obtained. Such spectra are similar to the differential absorption spectra of Spink and Wyckoff (Fig. 1 of reference 3), who used a cell contsining solution in the sample beam and had a matched cell containing water in the reference beam. Spectrum A shows the difference in the OH stretching bands quite clearly, the negative band being a measure of the greater absorption of water than of methanol. The negative OH stretching band declined in intensity with decreasing methanol concentration, as did the negative water deformation band in the 1700-1600 cm-l region (A-E, Fig. 3). The stretching band changed also in contour, and the deformation band

Fig. 3. Ratioed spectra using water aa reference: A: pure methanol. B: 76% methanol. C: 6O%methanol. D: 26%methanol. E: lO%meth~ol. Ineachcase, the referencespectrum wae that of pure water; e.g., spectrum A is the result of the ratio of spectra B and D of Fig. 1.

shifted in frequency, being at 1640 cm-l for pure water, 10% and 25% methanol solutions, near 1650 cm-l for 60% methanol solution, and near 1660 cm-l for 76% methanol solution. Such changes can be clearly seen in scaJe-expanded segments of spectra such as Fig. 4. However, the 1460 cm-l OH deformstion band of methanol did not shift, and there were no changes in the positions of the other methanol bands, when the methanol concentration was varied. The changes in the negative OH stretching band merely indicate the net changes in absorption; both the absorption caused by water and the absorption caused by

Infrared internal reflectionape&a of methanol-water mixturea

1791

Fig. 4. Water deformation bands: Sale-expanded segments of spectra of water (A) and of 60% methanol solution (B). The epectrum of methanol we used es tiference. The ordinates were expanded so that the two bands had the same inteneitiee.

the methanol hydroxyls might ohange. However, the change in frequency of the water deformation band (which implies that some change occurred in the intermolecular bonding of the water), and the constancy of the frequency of the methanol hydroxyl deformation band (which implies that the strength of the intermolecular bonding was not aBected), suggest that the changes of the contour of the negative OH stretching band were brought about mainly by changes of the absorption of the water component of the solutions. If the contour of the methanol band is unaffected, it would be more appropriate to compare the spectra of solutions with the spectrum of pure methanol, as in trace D of Fig. 2, which shows the OH stretching absorption as a pair of positive bands. In contrast, Malone and Flournoy obtained the intensities of the water OH and methanol OH absorptiona (both were taken to occur at 3360 cm-l) by a simple subtraction procedure; for each wavelength, the absorbance due to water was subtracted from the total to give that due to methanol alone. Scale-expanded segments showing the OH stretching regions of spectra such as D of Fig. 2 are collected in Fig. 6. The segments were scale-expanded to identical maximum and minimum ordinate values in order to bring out the changes in band contours more clearly. There were slight differenoes among the spectra of the 0, 10 and 26% solutions, and somewhat larger ohanges with the 60 and 75% solutions. The present results thus show that interactions between water and methanol did not affeot the methanol bands, so that these can be used for quantitative purposes, but that the water stretching and deformation bands were changed, so that these should not be used. It seems probable that similar effects would occur with other alcohol-water mixtures.

1792

R. T. YHU and M. J. D. Low

h

I 3700

,

I

3200 cni’

Fig. 6. Spectra of methanol solutions: The ooncentrationof methanol was: A, 0; B, 10; C, 25; D, 50; E, 76 vol.%, The epectrum of pure methanol was used as reference.

Glew htcs been able to measure the infi-ared spectra of dilute water in various solvents. With concentrated water, however, the intense absorption of water and the overlapping of the bands of water and solvent made the observations and interpretations d.i&ult [lo]. It seems worth noting that the present teohniques enable water to be examined at any concentration. Aoknowkdgmt-Support by Contract N00014-67-A-0467-0023 Research ia acknowledged. [lo] D. N. Gmw, in Hydrogen-Bonded Solvent System, and JONES,P.), Taylor & Francis, London, 1968.

from the Office of Naval

p. 133. (edited by

COVINGTON A. K.